Foreword. Stockholm, December Margareta Winberg, Minister for Agriculture, Food and Fisheries

Transcription

1 Foreword In March 1998 the European Commission presented its strategy for the enlargement of the EU, proposals for reforms of central policy areas and the financial perspectives Agenda An important section of this report deals with the EU's Common Agricultural Policy (CAP) and changes that will need to be made. As part of the work to analyse the consequences of the proposals in Agenda 2000 a network of experts on national agricultural policy has been attached to the Ministry of Agriculture, Food and Fisheries. The results of their work have made a valuable contribution to broadening and deepening the analyses of the Commission's proposals. The findings and analyses of the network have been important in helping us to understand the necessary changes to the Common Agricultural Policy that are imminent. Therefore, there is good reason to make their findings available to a wider circle of readers. The authors themselves are responsible for the content of the report. On 1 July 1999 an institute for agricultural economics is to be set up. This will have the main responsibility for economic analysis and investigation in the agricultural policy field, thereby giving the Ministry also in the future access to the analysis and investigative resources required for the long-term reform work. Stockholm, December 1998 Margareta Winberg, Minister for Agriculture, Food and Fisheries

2 Preface This report is a result of a co-operation between several institutions and organisations: Swedish University of Agricultural Science (SUAS), a consulting company Lantbrukekonomen, Uppsala University (UU), Swedish Board of Agriculture (SBA), Livsmedelsekonomiska Samarbetsnämnden (LES) and Swedish Environmental Protection Board (SEPB). The activities of the network have been co-ordinated by Ewa Rabinowicz. Peter Frykblom has functioned as secretary of the network until July 1998, followed by Erik Fahlbeck. Several persons contributed to the final product, which should be seen as a result of a joint effort of the participants in the network. Hence, it is not easy to allocate the responsibilities for different chapters in a fair way. A major contribution has been made by Lars Jonasson (Lantbruksekonomen) who performed the simulations that constitute the backbone of our analysis, especially in the first part of the report. These simulations are based on an agricultural sector model (SASM) that has been developed by Lars. He has also prepared background material for large parts of the part A and chapters 8, 9, 10 and, together with Bo Norell (SBA), for chapter 12. Ewa Rabinowicz (SUAS) has written chapters 7, 9, 12, 15, 16 and 17. She has also been responsible for editing of the report together with Olof Bolin, (SUAS) Erik Fahlbeck, (SUAS) and Yngve Andersson (UU). Erik has prepared chapters 1, 2, 3, 5, and 6. Chapter 13 has been written jointly by Peter Frykblom and Erik Fahlbeck (SUAS). Bengt Rundkvist (SEPB) has provided valuable comments to this chapter and to the report in general. Olof Bolin (SUAS) has prepared chapters 8, 10 and 11. Yngve Andersson has written chapter 14 and Per Persson (LES) prepared chapter 4. We have also benefited from valuable comments from the staff of the Department of Agriculture. The remaining errors are our own. December 1998 Ewa Rabinowicz

3 Ds 1998: 70 Summary and Conclusions i 1 Introduction Agenda Objectives of the study Swedish agriculture Methodology Impacts on Swedish agriculture at the sector level Present policy scenario Agenda 2000 base scenario Model results Impact on production and land use The impact on profitability and farm incomes Some impact related to the environment Impact on the budget and social welfare Summary and conclusions Impacts on farm level Background Some characteristics about the Swedish book-keeping survey for farmers Results Summary and conclusions Impacts of other reform proposals The French scenario The London scenario

4 Ds 1998: Summary and conclusions Sensitivity analyses Introduction Set aside and national envelopes Interpretation of the sensitivity analyses Lower price reduction on milk Lower price reduction on grains Lower price reduction on beef Lower price elasticities on demand Lower machinery costs Summary and conclusions Comparisons with other studies Differences in modelling approaches Differences in policy assumptions for the reference run Policy assumptions for Agenda Impact on use of acreage Impact on production Impact on farm incomes Summary Set-aside requirements Model results Mechanisms responsible for increased attractiveness of voluntary set aside and possible countervailing forces Regional implications Implications of shrinking ley area for tenancy Concluding discussion

5 Ds 1998: 70 9 National envelopes Impact on the budget and net contribution Distribution of the beef envelope Model-based analysis Is the model an adequate representation of reality? Summary Distribution of additional milk quotas Quota distribution Swedish consequences Production quotas become subsidy rights Summary and conclusions Principles of compensation Why compensate? Impact on Sweden Discussion Principles of modulation Introduction Income objective and fair income Possible definitions of fair income Equal pay for equal work Market related remuneration Modulation exercises with respect to income distribution Modulation and employment

6 Ds 1998: Reallocation of funds to rural development Summary and conclusions Some environmental impacts of the Agenda 2000 proposal Introduction Environmental policy objectives Land use Environmental programs Summary and conclusions Impact on consumers and taxpayers Total welfare Beef from steers or bulls? Market or regulation in an efficient provision of environmental benefits? Are system costs negligible? Summary and conclusions Rural development regulations and theoretical foundations of rural development policies Changes proposed in Agenda Conceptual issues Rural development - in search of a theoretical framework Development of rural regions in the future Policy implications and reforms of the structural funds Rural and structural policies in Sweden

7 Ds 1998: Future changes in regulations from a Swedish perspective Summary and conclusions WTO aspects Introduction Agenda 2000 and the URAA Is further liberalisation of agricultural trade likely? Key issues for the future WTO negotiations Expectations and credibility - are decoupled payments possible? Are decoupled payments desirable? Summary Eastern enlargement Introduction Agriculture in the CEECs - problems and potentials Development of agricultural policies in the CEECs- a short overview Impact of Agenda 2000 through Europe Agreements Eastern enlargement and the unreformed CAP CAP and CEECs Impact of the Agenda 2000 proposal Problems with the Agenda 2000 proposal Long-term options Is re-nationalisation an efficient long term solution? Summary and conclusions

8 Ds 1998: 70 Summary and conclusions i Summary and conclusions The objective of this study is to analyse the effects that could follow from an implementation of the Agenda 2000 proposal, focusing mainly on implications for agriculture in Sweden. Results in the study are derived from two kinds of methodology; model-based calculations and principally oriented reasoning about specific issues related to the CAP and the Agenda 2000 proposal. The impact on Sweden has been assessed using an agricultural sector model, SASM, developed by Lars Jonasson. SASM is a regionalised programming model based on farm accounting data. The impact of Agenda 2000 has been evaluated by comparing the expected effects of this reform proposal with a scenario representing a continuation of present policy. This "present policy" scenario represents a long-run, optimal adjustment to the current level of prices, direct payments, quotas etc. assuming that the agromonetary system will be abolished. Consequently, the outcome of present policy differs from preliminary figures for The calculations are based on the assumption that agricultural commodity prices will fall as deep as the proposed cuts in intervention prices. Within the Agenda 2000 scenario the so called national envelopes have been designed in such a way as to favour extensive forms of production. National supports to milk production in northern Sweden have been kept unchanged, implying lower support per cow, as total production will increase in response to expanded allocation of quotas to this region. The study is divided in two parts, A and B. The first focus on implications for Swedish agriculture and builds mainly on calculations with the chosen national model. The second part addresses a number of questions related to the Agenda 2000 proposal. Part A The overall picture of the estimated impacts of an implementation of the Agenda 2000 proposal in Sweden indicates limited changes on agricultural production. The model used identifies only minor economic consequences in real terms.

9 ii Summary and conclusions Ds 1998: 70 Table i Production under different scenarios (1000 tons) 1998 preliminary Present policy Agenda 2000 base Change in per cent* Grains 6,462 5,173 4,728-9 Oil-seeds Sugar beets 2,801 2,651 2,651 0 Roughage/pasture 4,413 4,396 0 Milk 3,280 3,300 3,373 2 Beef preliminary represents average yields at actual area for crops and animals as represented by 1997 actual figures. * Changes in per cent between the present policy scenario and the Agenda 2000 bases scenario Source: Statistics Sweden (SCB) and model estimations in SASM. The main reason for the modest effect on production is that agricultural production is strongly regulated by quantitative policy measures. Such measures limit the possibility to reallocate production and thereby reduce real economic impacts. As long as all the existing instruments that regulate production, i.e. producer rights, like animal headage premiums, direct payments to arable land etc., are kept unchanged within the CAP, large parts of the economic consequences that may follow the reform proposal, tend to end up as changes in values of quotas, producer rights and land. To attain substantial changes in production behaviour, on behalf of the farmers, a more far-reaching reform than the Agenda 2000 proposal of the CAP seems to be needed. Model analyses also illustrate the idea that, under such a regulated system as the CAP, many of the identified changes become strongly related to the technological requirements in the different regimes, or more generally: "the devil is in the details". Consequently, results are strongly related to the chosen technology, the chosen political requirements and to the physical relations in agricultural production. Changes in world market prices seems to be of minor importance. Moreover, since the Agenda 2000 proposal opens up for national incentives in many areas, but sticks to the production restriction instruments, the final outcome according to agricultural production will, within the national quota systems, become strongly related to specific Swedish policy decisions. The result is similar to results obtained by other studies. The overall impact of Agenda 2000 on prices, quantities and incomes in European agriculture appears to be relatively small according to available studies. Generally speaking, most of the impact, as predicted by those studies, is not due to changes of profitability but to changes in supply management parameters. Expansion of production of milk, cereals and oilseeds is due to

10 Ds 1998: 70 Summary and conclusions iii the changing of supply restrictions on those commodities. With respect to grain and beef production, Swedish results seem to differ considerably from calculations conducted on the EU level. Production of those commodities is expected to decline in Sweden but expand in the EU. To some extent this is a result of differences in assumptions and methodological approaches. The chosen model takes into account differences in land quality and allows a choice to produce or to set aside arable land. Thereby, a more realistic representation of the opportunities facing the producers is obtained. Similarly, the de facto decoupling of the suckler cow premium, which is not incorporated in other models, contributes to the decline of beef production, estimated for Sweden. Sensitivity analyses conducted indicate that results are robust under reasonable variations of assumptions. Despite the relative small impact on production, existing farmers will, however, suffer from lower land values, but will also realise increased production quota values. The estimates show that the value of the milk quotas will increase, due to lower costs for feed grains, silage and pasture. There will also take place a capitalisation of bovine premiums in mother animals. Costs for silage and pasture will decrease as a result of lower costs in land use. Model calculations also identify increased values of the implicit producer rights in meat, i.e. headage premiums for suckler cows. These premiums will also increase in value due to raised bovine premiums. Also sugar quotas show an estimated increased value, even though sugar isn t included in the reform proposal. Sugar quotas will be more valuable since the cost of land in alternative use decreases. As a consequence of marginal impacts on production and consumption, estimated total welfare changes become limited. For Sweden as a whole it is possible, and even probable, that the net and short run economic consequences end up in red. The Agenda 2000 reform proposal will, however, lead to redistribution; with losses for tax-payers (partly) compensating loosing farmers, while consumers will become gainers. Since Sweden is a net contributor to the EU-budget, it may be the case that the increased burden on Swedish tax-payers will offset the estimated food sector net welfare gains in a national general equilibrium context. As long as the battery of quotas, premiums, producer rights etc. are kept, it is reasonable to believe that the administrative burden will increase. If crosscompliances, direct payments to virtual cows, modulations, national envelopes etc. are implemented, the necessary increase in administrational efforts may be high enough in themselves to offset the otherwise estimated minor, but positive, welfare effects.

11 iv Summary and conclusions Ds 1998: 70 Part B A specific national set-aside policy seems necessary, since the upper limit of set-asides is crucial according to a number of important issues such as: the use and non-use of arable land, the price of grains, the self-sufficiency of grains, environmental consequences and to the compensational outcome of the Agenda 2000 proposal. If no upper limit is chosen it seems profitable for farmers to set aside the main part of their arable land, as they thereby are able to totally avoid machinery costs, implying increased "arm-chair farming". Even if that, in some dimensions, may be environmentally beneficial, the legitimacy of a policy paying farmers for using their land for non-production may be severely threatened, especially as the main part of the domestic grain consumption in such a case must be based on imports. Set-aside payments may, at a first glance, be considered as the final solution to the problem how to decouple agricultural support. This may not really be true, though. First, if normal farming practices are required to be eligible for support, some production will be necessary, even if it without the direct payments would be non-profitable. Second, if the upper limit is below 100 per cent set-aside, farmers must keep some machinery to fulfil the cultivation demands and thereby produce more, compared to a situation with no machinery and no production. In both cases, the technical requests linked to the payment of set-aside premiums will, in the Swedish case, drastically enhance production. Interesting is also that the link between direct payments and crop production is enforced with the Agenda 2000 proposal, making the support still more coupled. In lowering grain prices direct payments become more decisive than previously in making production profitable, as a larger share of the crop production needs acreage payments necessary to pass the zero profit limit. The same tendency can be seen in animal production and a general observation is therefore that with reduced intervention prices and lower market prices the increased direct payments tend to become more production stimulating. In other words, one of the basic ideas with direct payments, i.e. the decoupling of support to farmers, is partly lost when market prices become so low that production is not profitable without direct payments, and a "coupling effect" may well occur. Direct payments may be seen as partly compensating for reduced producer prices on the national, or sector, level. The compensation content is, however, weakened when the distributional content of compensation payments are considered, especially since the means are used for raised ambitions in environmental and regional policies. Also, the national quota expansion for Sweden has unexpected distributional implications, above all,

12 Ds 1998: 70 Summary and conclusions v it implies redistribution of production and wealth in milk production from high productivity agricultural areas to less productive ones. Focusing on the environmental aspects, an implication of the Agenda 2000 proposal, with an upper limit for set-aside at 50 per cent, indicates an increase in the use of pesticides. On the other hand, the estimated leakage of nitrogen in different forms, to air and water, is somewhat reduced, as well as the use of fuel. The total acreage of well-maintained grazing land is almost unaffected. These conclusions are valid for all analysed scenarios, with the exception for the case when no upper limit for set-aside land is determined. The choice of an upper limit for set-aside turns out to be the single most determining factor for the environmental effects. If no such upper limit is determined, all environmental effects are significantly stronger compared to the base scenario. In relation to regional differences in problems related to agriculture, not only environmental, model results indicate that both environmental and economic improvements are possible if such differences are considered when determining the policy. Compared with the original Agenda 2000 proposal the introduction of national envelopes implies an increased net burden for Sweden. The Swedish share of the estimated budget increase, that is necessary to finance the national envelopes, amounts to 21 million ECU. Livestock producers in Sweden would receive 13 million ECU, which results in a negative net balance of 8 million ECU. Policy discretion with respect to the use of envelopes is limited. The results are similar for different alternatives, but the underlying structure of incentives would be different. If payments are directed to pasture land, it will be profitable to keep such land, but the grazing animals as such will, however, not be profitable. Such a system creates an incentive to keep pasture land but requires supervision. If payment are directed to animals it will, on the other hand, not be profitable to graze natural pasture land. In order to stimulate permanent use of such land payments directed to animal production must be concentrated to extensive breeding forms. Model results point in the direction of a trade off between efficient production of ruminant meat and permanent use of natural pasture land. It also seems clear that the incentive structure of farmers tend to become even more complex than it already is. One of the most obvious examples is the beef production in Sweden. If the Agenda 2000 proposal is implemented, a beef-producing farmer will receive incomes from many sources: meat, headage premiums, extensification payments, national envelopes, compensation payments and a number of environmental programs. For the farmer it may become extremely complicated to identify what kind of production packages that maximises his total profits. This phenomenon will also make it difficult for administrators to influence production in a specific direction.

13 vi Summary and conclusions Ds 1998: 70 The crucial issue in the unfolding WTO negotiations will be the status of direct payments, especially whether such payments will be accepted and on what terms. A key question in this context is the impact of the direct payments on the incentives to produce, since decoupled payments can be considered as so called green box measures. American direct payments, (PFCs), are notified as green measures. A comparison between PFC payments and direct payments in the EU, under the assumption that 100 per cent set-aside is allowed, indicates that the PFC payment may be more production enhancing. If all direct payments are distributed to land and harvesting required, the differences between the Agenda 2000 base version and PFCs are small. Allowing for 100 per cent set-aside, makes EU payments more decoupled than PFC. Conditions attached to production obviously matter. However, even if payments would be formally decoupled such a policy may not be credible. The comparisons of alternative methods to design modulation of payments indicate that some of the transfers may be too high from the point of view of fairness, but it is difficult to design a modulation programme in a consistent and fair way. Without access to data on costs of production, other incomes and wealth, it cannot be known for sure that incomes are too high. If only the highest payments are reduced, the impact of modulation is almost negligible. A modulation of direct payments, rather than of all payments, appears arbitrary. The difficulties to design a reasonable modulation of payments reveals the fundamental dilemma of the CAP. The CAP originated from social concerns and has been focused on the farm income problem. The alleged low farm incomes have, though, been seen as a sector, and not a social, problem. The remedies have been designed accordingly and consisted of price support. The switch to direct support makes it technically possible to introduce modulation. Introducing some socially motivated restrictions on a sector policy will, however, not result in having a consistent social policy. It is generally agreed that rural development policies should play a more important role for the future vitality of rural regions. Changes that are proposed in Agenda 200 consist, however, mainly in some minor simplifications and reorganisations of existing measures. One of the explanations of depopulation of rural regions has been the result of declining share of food in peoples budget expenditures and low competitiveness of rural regions in producing those commodities that have been demanded instead. Rural development policies will hardly be able to counteract the continuous decline of agricultural employment. Instead, rural development policies should enhance competitiveness of rural regions by making some of the benefits that contribute to higher efficiency in urban regions available in rural areas and by removing negative implications to

14 Ds 1998: 70 Summary and conclusions vii such regions that may emerge as by-products of policies in other areas. Such benefits could include access to larger pools of resources. The principle of policy integration that will be followed, if Agenda 2000 is implemented, should be the ability to improve efficiency, by achieving economies of scope, avoidance of overlapping and by making it possible to handle larger projects. Closest potential for efficiency gains may be found between projects based on environmental support and rural development measures devoted to environment and small scale processing. Implementation of Agenda 2000 proposal contains some "good news" from the point of view of alleviating the enlargement of EU towards East and Central Europe. Prices for consumers would be lower, some simplifications of the CAP have been proposed, funds for rural development and restructuring, that the CEECs have been promised are, in the long run more efficient than boosting agricultural production in response to higher prices. The WTO commitments will also be facilitated, since the expansion of production at lower prices and without direct payments will be lower than otherwise. The long term problems of the CAP have, however, not been solved. Milk and sugar quotas are retained. Those are detrimental for the CEECs, which may end up absorbing structural surpluses from the incumbent members of the EU. The CAP is still highly complicated and the resulting structure of support to agriculture would, with Agenda 2000, become extremely unbalanced. Farmers in the CEECs would be forced to compete on uneven terms. End notes The existing CAP can be criticised for having fewer means than objectives. The reform proposals in Agenda 2000 can, on the other hand, be said to propose a broad range of instruments with conflicting ends. Above all, it is hard to distinguish the final result of the Agenda 2000 proposal according to the outspoken objective of enhancing the competitive advantage of EU agriculture on international markets. Is the Agenda 2000 proposal an improvement from the point of view of efficiency and equity? The major trust of the proposal is the replacement of price support by direct payments. It can be argued that such a change is a first step in an ongoing process of reform, but Agenda 2000 can hardly be seen as the final outcome. As long as the impact on production is modest, the related net welfare gain may be non-existent, if not negative, taking into account the administrative burden of distributing direct payments. If, after a shift to direct payments, the same producers produce the same commodities, using by and large a similar technology, price support may be more efficient. The

15 viii Summary and conclusions Ds 1998: 70 advantage of direct payments is that such payments make targeting, and hence a reduction of the total support, possible. The transparency of direct payments makes them open to public scrutiny. In the long run, the only permanent payments that can be justified from an economic point of view are those related to provision of public goods and positive external effects, most notably environmental benefits. In the short run equity considerations may be important, though, if the shift to a new system needs to be implemented gradually. Also the impact on consumer welfare is related to total transfers. If an unchanged burden of support to farmers is transferred from consumers to taxpayers not much is gained. Welfare gains on consumption will be more or less eliminated by distortions related to taxation since lump sum taxes only exist in economic textbooks. From the equity point of view, it should be observed that consumers and taxpayers strongly overlap. Also, with respect to other major challenges facing European agriculture, WTO and the Eastern enlargement, Agenda 2000 appears to be a start, but not a final solution. Excluding the new CEEC members from direct payments is not a viable long-term option. Perhaps, the Agenda 2000 proposal is not even possible as a short-term relief. Eventually, direct payments have to be extended to the new members or reformed. Long term challenge for European agriculture, in the context of international trade, is not to find new ways of paying the same amount of subsidies, but to identify what kind of payments that are legitimate to protect the European environment and to develop the countryside.

16 Ds 1998: 70 Introduction 1 Part A: Modelling the impacts of Agenda Introduction Reform proposals under Agenda 2000 have a number of goals such as increasing the competitiveness of the agricultural production within EU, easing an enlargement of the Union towards the East, directing production towards more ecologically sustainable approaches and to make the CAP more accepted among citizens. Integration is one of the most important issues in Europe. For the European Union (EU), the Common Agricultural Policy (the CAP) is one central area in the integration process. In 1997 the European Commission presented a reform proposal covering the CAP and the EU structural funds. The CAP has been questioned for many years. A number of weaknesses have been stressed and the present CAP seems to be costly to producers, consumers and taxpayers. Expenditures related to agriculture still cover almost 50 per cent of the total EU budget. After the MacSharryreform in 1992, factors like internal inconsistencies, complexity and budget pressures motivate further CAP reforms. For the enlargement of the EU towards East and Central Europe, the CAP is of ultimate importance. Further, it seems reasonable to believe that the next WTO round will continue the work started under the previous Uruguay round. For national governments, as well as for the EC-Commission, a comprehensive analysis seems required to elucidate the consequences of the reform proposals made in Agenda 2000 according to the agricultural policy. Will the proposals made meet the anticipated challenges? Will it

17 2 Introduction Ds 1998: 70 ease a further integration of Europe, an enlargement towards the East and the coming WTO negotiations? And if so, will the results be acceptable from a welfare economic point-of-view? 1.1 Agenda 2000 Agenda 2000 prolongs the strategy of the 1992 reform, i.e. a further shift from price support to direct income support. That approach is proposed, in particular, for cereals and beef. The level of the cereal intervention price is proposed to be fixed in one step at ECU per ton, i.e. 20 per cent below its present level. A non-crop specific area payment is proposed at ECU 66 per ton multiplied by the regional cereal reference yields of the 1992 reform. This payment may be lowered if market prices are sustained at a higher level than currently foreseen. For beef, the market support will be reduced in steps, totally by 30 per cent. The intervention system for beef will be abolished and a private storage regime introduced. Existing headage premiums will be increased and a new headage payment system for dairy cows introduced. Supply management measures are still provided, but they are intended to play a minor role. The main objective of the proposals is to improve the price competitiveness in the sectors. In the case of milk, the supply management system will be prolonged to year 2006 and will continue to play a major role for the dairy sector. For milk, cuts in intervention price of 15 per cent are proposed. An annual headage payment for dairy cows will partly compensate for this. Generally, compensational payments proposed are not intended to fully compensate for price reductions. No degressivity and no time limits are proposed for the direct payments. In addition, individual ceilings on the total amount of direct payments under different market regimes, i.e. modulations, are proposed. The Agenda 2000 offers Member States the possibility of modulating direct payments in all agricultural sectors on the basis of labour input used on the farm as well as the possibility of introduction of cross-compliances. The details are summarised below.

18 Ds 1998: 70 Introduction 3 Agenda Summary of proposals Cereals: reduction of intervention price from ECU/t to ECU/t, increase in indirect payments (non crop specific and applicable to cereals, oilseeds, non-textile linseed and set-aside) per hectare from 54 ECU/t to 66 ECU/t. Reference rate for compulsory set-aside will be fixed at 0 per cent. Beef: reduction of effective market support from 2780 ECU/t to 1950 ECU/t. Replacing intervention by private storage. Increase in direct payments for suckler cows from 145 to 180 ECU + additional payment for bulls from 135 to 220 ECU + additional payment for steers from 109 to 170 ECU + additional payment, new payment for cows 35 ECU. Payments for cows are annual, one payment per bull, two payments for steers. Additional payments are based on the envelope model (hectare or headage). Milk: The quota regime is extended to Additional quotas are introduced for young farmers (1 per cent) and mountainous areas and Nordic zones (1 per cent). Support prices are to be reduced by 15 per cent, compensated new direct payments for dairy cows amounting to 100 ECU per unit + additional payments (headage or per hectare) according to national preferences (envelope model). Horizontal provisions: Direct payments above are reduced by 20 per cent, payments above are reduced by 25 per cent. Cross compliances are left to the discretion of the Member States who may decide on appropriate measures. Agri-environmental measures: Funds made available from cross compliances or modulations remain available for MS as additional environmental support. LFA-payments based on per hectare basis. Rural development: RD1(agro-environment, afforestation, early retirement and LFA), RD2 (investment in agricultural holdings, marketing, forestry, development of rural areas). Financed by EAGGF Guarantee except for RD2 in objective 1 regions.

19 4 Introduction Ds 1998: Objectives of the study The study aims at providing an analysis of possible impacts of the Agenda 2000 proposal on Swedish agriculture. The study focuses primarily on impacts on Sweden. In addition, some general issues, potentially relevant to the Union as a whole and indirectly affecting Sweden, are included. When information is available, comparisons with other Member States are made. The analysis centres on the impact on economic efficiency. However, equity considerations are also included in the analysis, e.g. in the case of modulation. The outline of the report is as follows. The report is divided into two main parts. First, an agricultural sector model is used to estimate impacts of the reform on production, consumption, governments expenditures, trade etc., in Sweden. The second part provides a discussion of specific issues related to Agenda 2000 and to the model-calculations, such as alternative designs of modulation, set-asides, national envelopes, etc. In the first part, the mathematical model is presented, followed by a comprehensive presentation of the base run scenario. The presentation includes impacts on production, land use, livestock numbers, direct payments, consumption, trade as well as welfare consequences for producers, consumers and tax-payers. Figures are shown for the national and in some cases for the regional levels. The base run scenario corresponds to proposals made from the Commission on March 18, 1998 and it is assumed that the Agenda 2000 proposal is fully implemented. In cases where the proposal leaves the design of the regulations to the discretion of the national governments, assumptions are based on recommendations made by the Swedish Ministry of Agriculture. The base run scenario is complemented by other policy scenarios advanced in connection with the Agenda 2000 proposal, namely the French proposal of cutting direct payments to be able to reallocate funds to rural development and a proposal for a radical reform of the milk regime advanced by the "London club". Since Agenda 2000 opens up for national discretion of the policy, two different scenarios concerning the use of envelopes and the set-aside requirements are considered. Impacts on farm level are also examined to complement the sector model used. Scenario designs are the same for the farm level. In the base run, full impact of proposed price reductions is assumed. Implications of alternative assumptions are examined in the form of

20 Ds 1998: 70 Introduction 5 sensitivity analyses. In addition, reliability of the results is discussed, focusing on key methodological assumptions in the model. Part two of the study is devoted to the discussion of special issues. Several of the new regulations, notably the national envelopes, may be implemented in different ways. Impacts of varying assumptions are examined separately, keeping the remaining factors constant. This applies to the set-aside requirements, the national envelopes and the distribution of additional milk quotas. The major feature of the reform proposal is to switch from price support to direct payments. As direct payments are intended to compensate for the price cuts, we discuss principles of compensation. A critical examination of the different methods and principles of modulation is carried out as well. An environmental impact assessment of the Agenda 2000 proposal is presented as well as a discussion that focuses on the impact on food consumers. Changes in horizontal regulations, as well as the theoretical foundations of rural development policies, are analysed. Comparisons with other member states, based on other studies, are also presented. Finally, a discussion of issues related to the next WTO round and the Eastern enlargement completes the analyses in part two of the report. 1.3 Swedish agriculture Before presenting the model and the results, a brief introduction to the agricultural sector in Sweden is given below. Primary production in agriculture is about 1.1 per cent of the Swedish GDP and lesss than 2 per cent of total employment (Statistics Sweden, SCB). When up- and downstream industries are added, figures grow substantially. Today, there are about farmers in Sweden, of whom may be characterised as full-time farmers, (i.e. they have labour requirements in agriculture of 1600 or more standard hours per year, figures from Statistics Sweden, SCB, 1998). More than 75 per cent of farmers' assessed income comes from non-farm sources (Statistics Sweden, SCB). On the plains in Southern Sweden, conditions for cropping are similar to those in Denmark and in the northern parts of Germany. Soils have a high clay content; the growing season starts in April and ends in September. In the southernmost parts sugar beets are grown as well as grains, oilseeds, potatoes and vegetables. Pig fattening and milk

21 6 Introduction Ds 1998: 70 production are also important in these areas. Grass can be harvested two to three times a year. The northern parts of Sweden, together with the highlands and the woodlands in southern Sweden, have less fertile soils and a shorter growing season. Conditions for agriculture are similar to those in Finland and Norway. In these areas, milk production is by far the most important agricultural commodity. Many farmers, that used to have milk production, have, in recent years, switched to beef production because of raised productivity in milk production in combination with the introduction of quotas and environmental programmes. The average size of a Swedish farm is today almost 32 hectares of arable land, but full-time family farms without animals may be hectares or larger. In counties with fertile arable land in the middle and southern parts of Sweden the average size of farms is between 40 and 60 hectares arable land per farm. The average herd size of a milk producer is today about 30 cows, while those farmers who build new stables may establish herds of more than cows. Agricultural production in Sweden has never been explicitly exportoriented. For decades, the official agricultural policy has focused on national self-sufficiency. When support programmes and border protection measures stimulated production above domestic consumption levels, exports took place with the help of export subsidies. In 1990, the Swedish parliament decided on a radical agricultural policy reform - an internal deregulation. The main idea was to maintain border protection and eliminate all internal market regulations. A substantial adjustment programme gave temporary compensation/support to farmers during a five-year period. National programmes, mainly linked to environmental aspects within agriculture, were important parts of the reform. The deregulation intentions never came to a fulfilment, since the Swedish government applied for EU-membership in References: Yearbook of Agricultural Statistics Statistics Sweden, SCB, Halmstad, Sweden.

22 Ds 1998: 70 Methodology 7 2 Methodology In analysing effects of the Agenda 2000 proposal a sector model of Swedish agriculture is used. The model is called SASM-95 and has been developed by Lars Jonasson. A comprehensive presentation of the model can be found in Jonasson (1996). In short, the model can be summarised in the following box: Characteristics of SASM-95: Sector model covering Swedish agriculture and primary processing. Mathematical programming model. Multi-regional model with 10 production regions, 5 market regions and transport activities. Comparative-static equilibrium model with some dynamics in fixed assets. Demand is represented by regional linear demand functions and export at given prices. Supply is derived from production activities within the model and by import at given prices, separate from export prices. Production technology is represented by detailed regional crop and livestock budgets connected with a set of biological, technical and political restrictions. The model maximises total welfare of producers and consumers. Consumer behaviour is represented through price elasticities on a number of food commodities. Consumer food prices are based on farm gate prices plus fixed and variable costs for transports and processing. Traditional assumptions for utility maximisation are used (i.e. consumers prefer more and cheaper food to less and more expensive food). Producers are assumed to be profit maximisers. In the model, farmers are represented by a number of production activities. Each activity represents production of a commodity. According to allocation

23 8 Methodology Ds 1998: 70 of arable land, the model can choose between wheat, rye, oats, barley, rape, sugar beets, potatoes, grass for forage, ley, natural pasture, fallow and "others". For each crop commodity, fixed regional input cost structures per hectare are determined, i.e. each activity is given a fixed per hectare use of capital, machinery, tractor hours, man hours, fertilisers, pesticides, etc. Each activity also has a regional given output per hectare and yields per hectare are the regional averages. The model is based on ten production regions in Sweden and each region has its own cost structure and yield. To some extent, actual differences in farm structures, natural conditions and transport costs etc. are represented in the model. In each region, the arable land is split into two groups, one consisting of "better land" and one of "worse land". Demanding crops such as sugar beets, bread grains, oilseeds and potatoes can only be grown at "better land". Grass for forage and feed grains can be grown on all kinds of land, but feed grains give a 10 per cent lower yield on "worse" land. The classification is based on observed growing patterns of farmers, i.e. in each model region the share of "better" land is determined by the share of production that can be grown only on more fertile land (sugar beets, bread grains, oilseeds and potatoes). Also milk, meat, pork, sheep, poultry and eggs can be produced in the model. With these activities follow fixed costs for input use and output productions per head that vary regionally in the same way as for the crops. Within this general structure of farm production, processing and consumption activities it is possible to incorporate a large number of political CAP instruments. Regulations, restrictions, payments, etc. that are determined per input, per output, per hectare or per animal are easily included. Technology is represented through the input and output cost structures, included in the crop and livestock budgets. The farm structure is embodied in the supply of inputs. Inputs can be variable, intermediate, fixed or quasi-fixed. Most inputs are variable, e.g. fertilisers, pesticides, fuel and labour. These inputs are available at a fixed (infinitely elastic) price. Intermediate inputs (products) are inputs that are produced in one farm activity and used in another, e.g. forage, pasture grass, piglets and calves. Here, supply and demand has to be equal and regional equilibrium prices are obtained. If prices differ between regions, interregional trade may occur. Fixed inputs are typically represented by land. In each region a specific acreage is available for agriculture. This acreage can be used or not, but it cannot be expanded. If all available land in a region can

24 Ds 1998: 70 Methodology 9 be used for profitable production, land is a limiting factor and a land rent occurs. This land rent is an important part of the calculated producer surplus. No capital costs connected with present land values are included in the model. Quasi fixed inputs are fixed in the short run but variable in the long run, e.g. buildings and machinery. The model is run to elucidate intermediate-term effects. Assumptions have been chosen to represent possibilities in what may be interpreted as a ten-year perspective. There are no changes in technology or inflation representing a sequential time dimension. Time consuming development is, however, modelled. Capital stocks, i.e. machinery and buildings, depreciate with a certain percentage annually. The analyses are supposed to illustrate changes in a ten-year perspective and depreciation rates have been chosen such that a certain amount of the capital stock must be reinvested, to ensure continued production at present level. The remaining part of the present buildings will, though, create a producer surplus, similar to the land rent, if used for profitable production. No historical costs for these remaining buildings are included in the model. In milk production, the model can choose between three options; small, medium or large herds and stables. Initially, the actual distribution of buildings and herd sizes determines the starting point, but in the intermediate run and under changed policy regimes the model will choose the optimal distribution of reinvestments in milk production regionally and in increasing herd sizes. It may be pointed out that the structural allocation in size and region is partially determined by the present structure. Prevailing buildings can be used for free, while new buildings of larger size or in other regions are costly. When it comes to animal breeding, the model has no restrictions, apart from those that follow from the regional setting, i.e. the model can choose to expand, e.g. beef production in one region, as if this region was "one large farm", that without costs can pool animals to feeding and grazing areas. The regional redistribution of production is, though, softened by the cost advantage for continued production in present buildings such as for dairy production. Farmers behaviour is represented by the model's choice of the most profitable set of production activities. Choices are, however, restricted in a number of ways. In each region, biological restrictions on crop rotations may limit the use of crops, and growth in livestock is partly limited by birth-rates, existing buildings and animal feed. Some interregional transportation may occur but, at national level, breeding of beef is limited by the number of calves, etc.

25 10 Methodology Ds 1998: 70 EU-regulations, such as quotas, also limit the choice of activities. In Sweden the government has chosen to divide the country in two regions concerning the distribution of milk quotas. Quotas may be traded intra-regionally at administratively set prices, but are not tradable between these two regions. In the model, milk quotas are freely traded within the two quota regions and the equilibrium price indicates a quota rent. In reality, milk quotas can be traded within each region, at an administratively set price. When it comes to the Swedish environmental programme and regional support, Sweden is split into eight regions which also are represented in the model. The remaining areas in the model are separated into two regions, which makes a total of ten production regions in the model. Prices are determined by the model as a set of equilibrium prices, but the magnitude for traded products is limited by import and export prices. Export and import prices are exogeneously determined outside the model. They are set at levels that represent actual EU prices. For most products the Swedish export price equals the intervention price while the import prices equals the intervention price plus transportation costs to Sweden. It is assumed that farmers are unable to discriminate between domestic and foreign markets. Consequently, farmers are modelled as price takers. Even if the model assumes normal negative price elasticities for consumers, farmers do not act strategically, e.g., in order to restrict output for higher prices - with one exception. In setting the price of milk, it is assumed that a fixed part of the price for fresh milk is internally distributed to other milk-based commodities. This adjustment represents the old price-setting structure in farmers milkprocessing co-operatives. Even if there is no evidence for such a price setting mechanism today, it is notable that the price of fresh milk to consumers, as well as the farm gate price, is higher in Sweden than in most other EU-countries. Prices may vary within Sweden between, but not within, regions. Local differences in production and consumption lead to internal domestic trade in Sweden. Prices may differ according to domestic transport costs. The model is based on average data. Figures for yields and costs represent regional and group averages. Thereby, the model can represent the nation, or a region, in a simpler manner than a model based on non-linear cost and demand functions. As a result, model outcome becomes "threshold dependent", i.e. the model may over-react at certain critical levels. It may, e.g., be the case that at a certain price, large areas of arable land are used for grain production, but at a slightly